Permafrost has a significant impact on high latitude ecosystems and is spatially heterogeneous. However, only generalized maps of permafrost extent are available. Due to its impacts on carbon pools, subsurface hydrology, lake water levels, vegetation communities, and surface soil deformations, an understanding of spatial extents and depth of permafrost are critical for proper management and monitoring of these areas. In this study, we propose a method for accurately extrapolating Airborne Electromagnetic Resistivity (AEM) for regional permafrost mapping in the Yukon Flats Ecoregion (YFE), Alaska, through the use of regression tree models. Electrical resistivity serves as the proxy for permafrost presence in the AEM extrapolation portion of this study, as electrical resistivity increases dramatically as soil freezes. This method uses resistivity values, and other relevant data to predict near surface (0-2.6m) electrical resistivity at a 30-m resolution within the YFE. We also propose a piecewise regression model (Cubist) and a Presence/Absence active layer decision tree classification (See5) that use in-situ data and other relevant spatial data, to accurately estimate Active Layer Thickness (ALT) or thaw depth (0-122cm) at a 30-m resolution within the YFE.